The effect of local extended families in epidemics: the case of COVID-19 deaths in the US

Do extended family members that live near each other significantly affect the dynamics of large respiratory infectious disease outbreaks? While current epidemiological consensus recognizes household family as relevant, extended family (kin) groups are not viewed as playing an especial role, disregarding evidence of the frequent and unique support they provide each other. Indeed, if extended family was relevant, this disregard would likely lead to deficiencies in advice, modeling, and preparedness during an epidemic as neither the interaction patterns with extended family nor the effectiveness of non-pharmaceutical interventions (NPIs) on them are known. Here, through the creation of the first large-scale data set of obituaries in the US for research purposes ( ~ 1.4 million deaths), we show for the first time evidence that between the beginning of 2020 and end of 2022 (the height of the COVID-19 pandemic) there were surges of deaths among kin members living in the same cities, consistently spanning summer periods. The kin surges we observe started simultaneously with overall COVID-19 population surges but continued even as the overall surges subsided. These surge patterns are consistent with extended families maintaining contact even through strict periods of NPIs implementation, potentially making them an unrecognized reservoir of disease spread helping outbreaks persist. Further, we find indications that kin-related deaths are more prevalent among genetic kin and display gender heterogeneity. Our results call for a renewal on research on infection-relevant contact patterns among extended US families and on the effects that NPIs may have towards these contacts.